Critical assessment of CEO succession on organizational performance through descriptive and predictive research methods

Abstract

This research offers a comprehensive analysis of extant CEO succession literature to discover and illuminate previously unanalyzed variable relationships and potential areas for future research. An analysis is performed using both traditional regression-based statistical methods (descriptive) and machine learning methods (predictive) to compare the capabilities of the methodological approaches in analyzing complex relationships among the many variables in the underlying domain, ultimately demonstrating that the type of analytical lens used influences research outcomes. Results indicate that lead time is shown to be the greatest predictor of future firm financial performance following a change of CEO. The level of education of both the incumbent and incoming CEOs is shown to be virtually irrelevant to a company's financial performance when also factoring in several other relevant predictors. The downgrading of stocks by investment analysts is found not to impact firm performance. Previous CEO experience is determined not to be indicative of future company performance. Outgoing CEO age is demonstrated to be significantly positively related to future long-term firm financial performance. This research is the first to explore CEO succession and the impact of CEOs on firm financial performance using a machine learning methodology, providing methodological and theoretical contributions that should be considered by practitioners and researchers to advance research in these areas further.